Abstract

Adoptive transfer of T cells is frequently associated with unwanted side effects. In order to tackle these effects one could introduce a safety switch into the cells that permits their selective in vivo elimination. The human CD20 gene in combination with CD20 antibodies was recently proposed as a novel safety switch. In such a system, T cells may be genetically modified with a CD20-encoding vector prior to adoptive transfer. If necessary, CD20-transgenic cells can be eliminated in vivo through administration of CD20 antibodies, such as the chimeric antibody rituximab (RTX) that is currently used to treat CD20+ lymphoma. RTX activates the complement system and recruits immune effector cells, resulting in rapid death of CD20+ cells. Recently, a novel human CD20 antibody, Humab 7D8, was shown to have superior activity over RTX.

In this study a set of CD20-encoding retroviral vectors was generated, which either lacked or contained one or both of two regulatory elements:

  1. the woodchuck posttranscriptional regulatory element (WPRE) to increase CD20 expression, and

  2. the chicken hypersensitivity site 4 insulator element (INS) to achieve a position independent expression of CD20 and to increase the safety profile of the vector by preventing activation of cellular (onco)genes by the retroviral enhancer.

We found that the level of CD20 expression obtained with vectors containing INS was 2-fold lower than with vectors lacking INS. Additional inclusion of WPRE restored the level to that of the vector without INS. In addition, INS greatly enhanced the homogeneity of CD20 expression in T cells. Moreover, after 3 months in culture, all cells generated with CD20-INS had retained CD20 expression, while 60% of cells transduced with the control CD20 vector had lost CD20 expression.

Complement dependent cell kill (CDC) of both RTX and HuMab 7D8 was dependent on the level of CD20 expression (p<0.01). However, while very low CD20-expressing cells were completely resistant against RTX they could be effectively killed by HuMab 7D8. For maximal kill of CD20-high cells, a 100-fold lower dose of HuMab 7D8 was required, compared to RTX. In vivo efficacy was studied through bioluminescent imaging of luciferase+ CD20-transgenic T cells. After transfer of CD20+ cells in immune deficient RAG2−/−gamma c−/− mice, both CD20 antibodies were capable of eliminating >99% of CD20+ cells, prolonging survival of mice from 20 till 42 days.

In conclusion, we developed a safe vector that leads to homogeneous and stable expression of CD20 on human T cells. These cells can be killed effectively in vivo with HuMab 7D8, a recently developed CD20 antibody. This system will be applicable to other approaches that require inclusion of a safety switch in ex vivo modified cells.

Disclosures: Wendy Mackus, Paul Parren and Jan van de Winkel are employed by Genmab B.V.

Author notes

*

Corresponding author